Question & Answer
How are Enterprise IDs assigned ( EIDs )?
During standard processing of data within IBM® Initiate™ software, Enterprise IDs ( EIDs ) are assigned based upon various rules within the software and as dictated by the customer algorithm.
When a member is created, it is assigned an internal member record number (memrecno) and an Enterprise ID (entrecno). During processing, member records are linked to entities with other records. As members are added/updated, linked/unlinked, merged/unmerged or removed from the system, the collection of member records which comprise each entity can change which may result in the change of a member's and/or entity’s Enterprise ID.
In earlier versions of the IBM® Initiate™ software, the initial Enterprise ID assigned to a member was a number equal to their Member Record Number or memrecno. After the initial assignment and as entities changed, Enterprise ID assignment used a process of selecting an Enterprise ID from the entity set that was either the lowest number in the entity or the Enterprise ID that was most used in the set. The over-arching goal was to have the least number of Enterprise ID changes as possible, however the possibility and likelihood exists that any Enterprise ID associated with any individual member or entity could change.
As data sets continue grow and change, and as more customers elect to consume Enterprise IDs in their downstream systems, the memrecno=entrecno assignment method has proven to be more complex than desired. A scenario that illustrates this complexity is when a record unlinks from an entity and happens to be the base or “glue” member (i.e. the memrecno equals the entrecno, which also equals the entity’s entrecno). The unlinking member will maintain the Enterprise ID, while all of the records remaining in the entity would have their Enterprise IDs changed. Thus, a single change can trigger multiple EID changes.
In v8.5 and newer, entity identifier creation and modification has been simplified such that Enterprise IDs no longer follow the memrecno=entrecno assignment concept, but rather are assigned in sequence via the mpi_seqgen table when being added, linked/unlinked, or merged/unmerged. This method of assignment successfully maintains the “least number of changes” logic. With this modification, there is no longer any correlation between memrecno and entrecno. This means that "rules" or assumptions commonly understood in previous releases are no longer valid, such as:
- one record within a entity would always have a memrecno=entrecno assignment
- a memrecno=entrecno assignment always signified a singleton entity.
Was this topic helpful?
16 June 2018